Let’s say you have 100 people.
They can break into groups of any size.
Each group gets 100 toothpicks.
The goal: build the tallest structure with the toothpicks.
What is the optimal group size?
Maybe it’s a group of 100. Everyone works together contributing their best ideas and providing feedback on other ideas. Everything is shared and communal. Sounds nice, but seems pretty unlikely this will yield the tallest structure. Just logistics alone, there aren’t enough toothpicks for everyone to be trying ideas. Human nature is such that a lot of the 100 probably won’t say or contribute anything at all, and some loudmouth will assume control. Also: too many cooks in the kitchen.
On the reverse, is it groups of one? You’ll get 100 ideas this way. Everyone will try something. Maybe someone will stumble upon a really excellent idea. This seems better but also unlikely to be the best. There will be less feedback and idea iteration this way. Some people have the best ideas when their brain is seeded with existing ideas, and you don’t get that here.
So if it’s not 1 or 100, what is it? Two, five, ten?
It’s fairly easy to Google and find answers like 5-6, 4-8, 5-9, etc. I can’t promise you that’s the result of sound research, but with gut instinct and life experience, those numbers feel ballpark right.
I think of this in terms of business sometimes. Right now, Google is proudly promoting their Gemini AI thing, and comparing it to OpenAI’s existing powerful AI tools. Part of my brain goes: wouldn’t y’all be able to build more powerful things if you worked together? Whether that’s good for the world, I can’t say, but you’d think the models would be better, the computer power could be more efficiently used, the best ideas could be shared, etc. But then another part of my brain goes: maybe not. Too many cooks in the kitchen. Plus competition breeds innovation.
I think about it with browsers too. Wouldn’t the best browser be one where everybody who works on anything browser related worked together? Seems like yes. But also seems like no, because one organism doesn’t make for a healthy ecosystem.
I think it also hinges on what the goal is — building the tallest toothpick structure is a clearly defined goal that is trivial to measure. On the other hand Google and MS seem to have different goals here (let’s assume they share the long-term goal of maximizing shareholder returns. And in the very-short-term they’re both in the fiddle-around-see-what-we-can-build phase).
Ultimately Google wants to be your sole source of knowledge, and MS wants to be your sole operating system for life and work. That’s how each one of them knows to turn users into revenue. These goals are not that different overall, but they do dictate different approaches and trade-offs.
That’s also why we have different browsers — Google wants to maximize your content-to-ads ratio, MS wants you to stay inside their system, Apple wants to sell you hardware, and Mozilla… are trying to figure out if they still want to exist, I guess?
I think it could be nice to live in a world with a single, non-profit. open-source browser engine maintained by the W3C or the like (they don’t seem to need competition to keep up innovation). At the very least there would be some agreement on what the goal is. WinterCG is a step in the right direction of making the products more compatible, but I don’t think it makes these corporations’ goals more compatible.